BUG: Segfault due to float_precision='round_trip'

`round_trip` calls back into Python, so the GIL must be held.  It also
fails to silence the Python exception, leading to spurious errors.
Closes pandas-dev#15140.

Author: Phil Ruffwind <rf@rufflewind.com>

Closes pandas-dev#15148 from Rufflewind/master and squashes the following commits:

c513d2e [Phil Ruffwind] BUG: Segfault due to float_precision='round_trip'

doc/source/whatsnew/v0.20.0.txt

@@ -440,3 +440,5 @@ Bug Fixes
 
 - Bug in ``Series.dt.round`` inconsistent behaviour on NAT's with different arguments (:issue:`14940`)
 - Bug in ``.read_json()`` for Python 2 where ``lines=True`` and contents contain non-ascii unicode characters (:issue:`15132`)
+
+- Bug in ``pd.read_csv()`` with ``float_precision='round_trip'`` which caused a segfault when a text entry is parsed (:issue:`15140`)

pandas/io/tests/parser/c_parser_only.py

@@ -388,3 +388,10 @@ def test_read_nrows_large(self):
         df = self.read_csv(StringIO(test_input), sep='\t', nrows=1010)
 
         self.assertTrue(df.size == 1010 * 10)
+
+    def test_float_precision_round_trip_with_text(self):
+        # gh-15140 - This should not segfault on Python 2.7+
+        df = self.read_csv(StringIO('a'),
+                           float_precision='round_trip',
+                           header=None)
+        tm.assert_frame_equal(df, DataFrame({0: ['a']}))

pandas/parser.pyx

@@ -181,7 +181,11 @@ cdef extern from "parser/tokenizer.h":
         PyObject *skipfunc
         int64_t skip_first_N_rows
         int skipfooter
-        double (*converter)(const char *, char **, char, char, char, int) nogil
+        # pick one, depending on whether the converter requires GIL
+        double (*double_converter_nogil)(const char *, char **,
+                                         char, char, char, int) nogil
+        double (*double_converter_withgil)(const char *, char **,
+                                           char, char, char, int)
 
         #  error handling
         char *warn_msg
@@ -482,11 +486,14 @@ cdef class TextReader:
 
         self.verbose = verbose
         self.low_memory = low_memory
-        self.parser.converter = xstrtod
+        self.parser.double_converter_nogil = xstrtod
+        self.parser.double_converter_withgil = NULL
         if float_precision == 'high':
-            self.parser.converter = precise_xstrtod
-        elif float_precision == 'round_trip':
-            self.parser.converter = round_trip
+            self.parser.double_converter_nogil = precise_xstrtod
+            self.parser.double_converter_withgil = NULL
+        elif float_precision == 'round_trip':  # avoid gh-15140
+            self.parser.double_converter_nogil = NULL
+            self.parser.double_converter_withgil = round_trip
 
         # encoding
         if encoding is not None:
@@ -1699,17 +1706,31 @@ cdef _try_double(parser_t *parser, int col, int line_start, int line_end,
     result = np.empty(lines, dtype=np.float64)
     data = <double *> result.data
     na_fset = kset_float64_from_list(na_flist)
-    with nogil:
-        error = _try_double_nogil(parser, col, line_start, line_end,
+    if parser.double_converter_nogil != NULL:  # if it can run without the GIL
+        with nogil:
+            error = _try_double_nogil(parser, parser.double_converter_nogil,
+                                      col, line_start, line_end,
+                                      na_filter, na_hashset, use_na_flist,
+                                      na_fset, NA, data, &na_count)
+    else:
+        assert parser.double_converter_withgil != NULL
+        error = _try_double_nogil(parser,
+                                  <double (*)(const char *, char **,
+                                              char, char, char, int)
+                                  nogil>parser.double_converter_withgil,
+                                  col, line_start, line_end,
                                   na_filter, na_hashset, use_na_flist,
                                   na_fset, NA, data, &na_count)
     kh_destroy_float64(na_fset)
     if error != 0:
         return None, None
     return result, na_count
 
-cdef inline int _try_double_nogil(parser_t *parser, int col,
-                                  int line_start, int line_end,
+cdef inline int _try_double_nogil(parser_t *parser,
+                                  double (*double_converter)(
+                                      const char *, char **, char,
+                                      char, char, int) nogil,
+                                  int col, int line_start, int line_end,
                                   bint na_filter, kh_str_t *na_hashset,
                                   bint use_na_flist,
                                   const kh_float64_t *na_flist,
@@ -1739,7 +1760,7 @@ cdef inline int _try_double_nogil(parser_t *parser, int col,
                 na_count[0] += 1
                 data[0] = NA
             else:
-                data[0] = parser.converter(word, &p_end, parser.decimal,
+                data[0] = double_converter(word, &p_end, parser.decimal,
                                            parser.sci, parser.thousands, 1)
                 if errno != 0 or p_end[0] or p_end == word:
                     if (strcasecmp(word, cinf) == 0 or
@@ -1760,7 +1781,7 @@ cdef inline int _try_double_nogil(parser_t *parser, int col,
     else:
         for i in range(lines):
             COLITER_NEXT(it, word)
-            data[0] = parser.converter(word, &p_end, parser.decimal,
+            data[0] = double_converter(word, &p_end, parser.decimal,
                                        parser.sci, parser.thousands, 1)
             if errno != 0 or p_end[0] or p_end == word:
                 if (strcasecmp(word, cinf) == 0 or

pandas/src/parser/tokenizer.c

@@ -1773,11 +1773,9 @@ double precise_xstrtod(const char *str, char **endptr, char decimal, char sci,
 
 double round_trip(const char *p, char **q, char decimal, char sci, char tsep,
                   int skip_trailing) {
-#if PY_VERSION_HEX >= 0x02070000
-    return PyOS_string_to_double(p, q, 0);
-#else
-    return strtod(p, q);
-#endif
+    double r = PyOS_string_to_double(p, q, 0);
+    PyErr_Clear();
+    return r;
 }
 
 // End of xstrtod code

pandas/src/parser/tokenizer.h

@@ -201,7 +201,11 @@ typedef struct parser_t {
     PyObject *skipfunc;
     int64_t skip_first_N_rows;
     int skip_footer;
-    double (*converter)(const char *, char **, char, char, char, int);
+    // pick one, depending on whether the converter requires GIL
+    double (*double_converter_nogil)(const char *, char **,
+                                     char, char, char, int);
+    double (*double_converter_withgil)(const char *, char **,
+                                       char, char, char, int);
 
     // error handling
     char *warn_msg;